prompt-ai

Prompt-AI Library

Prompt-AI is a powerful library designed to optimize AI-driven prompt handling and response generation using the Gemini API. By introducing structured database management and efficient embedding retrieval, Prompt-AI significantly enhances performance, reduces response times, and provides a seamless solution for integrating AI models into various applications.

Key Features

• Efficient Embedding Management: Prompt-AI stores pre-generated embeddings in a structured database, significantly reducing computational overhead and improving response times.
• Real-Time Updates: Manage datasets and dataframes efficiently, ensuring that embeddings are generated once and reused across multiple sessions.
• Performance and Scalability: The streamlined approach enhances performance and scalability, making Prompt-AI ideal for chatbots, recommendation systems, and other AI-powered tools.
• Versatile Integration: Seamlessly integrates with Node.js endpoint servers, bridging different technologies and workflows.

Installation

To install Prompt-AI, use pip:

pip install prompt-ai

To upgrade to latest version:

pip install prompt-ai --upgrade

---

After Installation follow these steps to use promp-ai

1. Generate an API Key

To begin, you’ll need to generate an API key. Follow the link below to generate your API key:

Generate API Key

Brief Summary of Gemini Model

The Gemini model is a powerful AI-driven model designed for generating contextually relevant responses to user prompts. Unlike traditional approaches where embeddings are generated on each run, Prompt-AI integrates a more efficient workflow by storing pre-generated embeddings in a NoSQL database. This allows for faster response times and reduces computational overhead, making it ideal for applications like chatbots, recommendation systems, and other AI-powered tools.

3. Setting up MongoDB (In later versions: SQL and Cloud database will be added)

1. Create a Database in Mongo Atlas or MongoDB Compass (Which you feel good).
2. Create collection and Documents.
3. Set the document in this structure:

{
"id": 1,
"title": "Gork vs Chat-gpt",
"text": "In the rapidly evolving landscape of artificial inte...",
}

4. Using Prompt-AI to manage prompts and generate response

Prompt-AI provides two core functions to help you manage prompts and generate responses:

1) configure(mongo_uri: string, db_name: string, collection_name: string, columns: array, API_KEY: string, embeddings: bool)

configure(mongo_uri, db_name, collection_name, columns, API_KEY, embeddings)

This function configures the connection to your MongoDB Atlas and sets up the necessary parameters for generating embeddings.

• mongo_uri: string
  This should contain your MongoDB Atlas connection string.

mongo_uri = 'mongoDB connection string'

• db_name: string
  The name of your MongoDB database.

db_name = 'Database Name'

• collection_name: string
  The name of the collection in your database where the data is stored.

collection_name = 'collection name'

• columns: array
  An array of strings, each representing a field name present in each document of the collection. The field which contains ANSWER data must be named with 'text'.

columns = ['id', 'title', 'text']

• API_KEY: string
  The API key generated in the first step.

API_KEY = 'key generated in first step'

• embeddings: bool
  A boolean flag indicating whether embeddings need to be created (true) or if they already exist (false).

embeddings = True or False

This function call will return datasets in form of tabular dataframe.

id          title                    text                   embeddings
1       "chat-gpt features"    "chat-gpt has..."        [4.0322, 2.3344, 1.09...]
2       "Gork vs Chat-gpt"     "Gork have plenty..."    [1.1702, 0.4184, 5.19...]

overall configure function will look like this...

uri = 'your connection string' 
db = 'database_name'
col = 'collection_name'
API_KEY = 'api_key generated in first step'
column = ['id', 'title', 'text']
embeddings = True
dataframe = configure(uri, db, col, API_KEY, column, embeddings)

2) generate(user_prompt, dataframe)

generate(user_prompt, dataframe)

This function processes the user’s prompt, interacts with the database, and returns the AI-generated response. dataframe will be used here inside generate() function.

@app.post("/api")
async def generate_response(request: PromptRequest):
    # Extract the prompt from the request body
    user_prompt = request.prompt

    # calling generate() function with prompt and dataframe as parameter

    response = generate(user_prompt, df)

    return {"response": response}

Now on making post request to this endpoint keep prompt: string inside body of the post request in JSON format. And for handling response you can have it in this way:

const response = await axios.post('endpoint_url/api', {prompt});
res = response.data.response;

Here I have used axios, but you can also use fetch api to make post request and fetch the response in the same way.

---

With Prompt-AI, you can efficiently manage AI-driven prompt handling, leveraging the Gemini model's capabilities with enhanced performance and scalability. Whether you’re building a chatbot, a recommendation system, or any other AI-powered application, Prompt-AI provides a streamlined and powerful solution.

Happy Coding :)

---